Long COVID — Pathophysiology Hypotheses

3. Long COVID — pathophysiology hypotheses

Long COVID (WHO: post-COVID-19 condition) is a multisystem condition without a single, proven cause. According to the WHO case definition, the symptoms usually appear from three months after a probable or confirmed SARS-CoV-2 infection, last at least two months and cannot be explained otherwise [17]. In Germany, the AWMF S1 guideline "Long/Post-COVID" (living guideline, reg. 020-027) and RKI information resources provide recommendations and orientation on diagnostics and care; it is a consensus-based S1 (not evidence-graded S3) guideline [15][16]. Research in 2023–2025 is converging on several mechanisms that probably reinforce one another. Important for affected people and those treating them: none of the following hypotheses is established as the sole cause, and none alone explains all cases. Long COVID is heterogeneous and presumably to be divided into subphenotypes; the major reviews stress that the mechanisms probably do not compete but are interconnected [1][2].

Viral persistence (reservoir hypothesis)

SARS-CoV-2 RNA, viral antigens and spike protein have been detected in tissues months after the infection – among others in the gut, lymphatic tissue and blood plasma [3][4]. Findings on accumulation along the skull-meninges-brain axis so far come mostly from animal models and post-mortem studies and are not yet established for living Long COVID patients [uncertain]. Persisting viral antigen is regarded as a plausible but unproven ongoing trigger of chronic inflammation and is the target of antiviral studies (e.g. nirmatrelvir/ritonavir [Paxlovid], monoclonal antibodies). These substances are approved for acute therapy; their use in Long COVID is experimental/off-label. The randomised trials published so far (among others with prolonged nirmatrelvir/ritonavir dosing) mostly showed no convincing benefit; a treatment effect is currently not proven and should be assumed only within controlled studies [3]. Safety note: no unsupervised or prolonged intake of antiviral medicines outside medical care – nirmatrelvir/ritonavir has clinically relevant drug interactions.

Immune dysregulation and complement activation

Cervia-Hasler et al. (Science 2024) showed, in active Long COVID, a persisting complement dysregulation with an elevated terminal complement complex and signs of thromboinflammation; in those who recovered the markers largely normalised – a promising candidate biomarker that, however, has not yet been broadly replicated independently and is not clinically validated [5]. Linked to this, an ongoing systemic inflammation is being discussed.

Autoantibodies and autoimmunity

Functional autoantibodies against G-protein-coupled receptors (adrenergic β1/β2, muscarinic M2/M3, angiotensin AT1/ETA receptors) have been repeatedly described. These have been linked to autonomic dysfunction and disturbed vascular regulation; molecular mimicry and IgG transfer models (transfer of patient IgG into mice) support a possible causal contribution [6][7]. Assessment: the evidence is heterogeneous – some well-controlled studies found no consistent elevation of these autoantibodies, and a causal contribution in humans is not conclusively proven [uncertain].

Reactivation of latent herpesviruses

A reactivation of latent viruses, above all the Epstein-Barr virus (EBV), has been observed in a proportion of Long COVID sufferers and described in cohort studies as a prognostic factor; HHV-6 and CMV are also being discussed [8]. Whether this is a cause, a cofactor or an accompanying phenomenon is open.

Microclots and endothelial dysfunction

Pretorius/Kell describe fibrinaloid, amyloid-containing microclots in platelet-poor plasma that may be resistant to fibrinolysis and could impair oxygen transport in capillaries [9][10]. These findings have so far not been independently replicated in a standardised way and are methodologically disputed and are therefore to be regarded as a hypothesis. In parallel there are signs of persisting endothelial damage (raised von Willebrand factor, soluble thrombomodulin) that could link microclot, complement and hypoperfusion findings. Safety note: anticoagulant treatment or a so-called "triple anticoagulation" (dual antiplatelet therapy plus anticoagulation) based solely on microclot tests is not evidence-based, carries substantial bleeding risks and is explicitly not recommended outside of studies. The same applies to blood-washing/apheresis offers against "microclots", whose benefit is not proven.

Autonomic dysfunction (POTS/orthostatic intolerance)

Orthostatic intolerance and postural tachycardia syndrome (POTS) occur more frequently in strongly symptomatic patients [11]. Discussed mechanisms are autoantibodies against adrenergic/muscarinic receptors, small-fibre neuropathy, hypovolaemia and deconditioning. Diagnostics and therapy (e.g. standing/tilt-table assessment, non-drug measures) should be carried out with medical guidance.

Mitochondrial/metabolic muscle dysfunction and PEM

A muscle biopsy study (Appelman et al., Nat Commun 2024; n=25 Long COVID vs. 21 controls) showed reduced oxidative phosphorylation and altered succinate dehydrogenase activity, amyloid-containing (extracellular) deposits and, after maximal exertion, focal necrosis – large necrotic fibre areas occurred in around 36 % of the Long COVID patients [12]. This is discussed as an objective correlate of post-exertional malaise (PEM) and, pathophysiologically, supports the pacing principle (exertion within one's individual energy limit) over a rigidly graded increase in activity (graded exercise therapy, GET). Safety note: in pronounced PEM, GET can worsen symptoms; current guidelines (among them NICE) recommend pacing and advise against forced increases in activity.

Gut-serotonin axis and neuroinflammation

Wong/Levy et al. (Cell 2023) postulate a cascade: viral persistence in the gut → interferon response → reduced tryptophan uptake → serotonin deficiency → disturbed vagal signalling → cognitive symptoms; in the animal model, serotonin precursors or an SSRI reversed symptoms [13]. This is a still largely preclinical hypothesis – a conclusion that SSRIs are a Long COVID therapy in humans cannot be derived from it. SSRIs are approved (for other indications); their use against core Long COVID symptoms would be off-label and unproven in efficacy; stopping or starting them on one's own should not happen without medical consultation. In cognitive impairment, signs of a disturbed blood-brain barrier and ongoing systemic inflammation have also been reported [14]; PET studies on microglial activation are contradictory, so that "brain fog" is regarded as multifactorial.

Overlap with ME/CFS and assessment

A relevant proportion of those affected meet, from about six months onwards, the diagnostic criteria for ME/CFS with the hallmark symptom PEM; several of the mechanisms named overlap, which suggests a (partly) shared pathophysiology [1]. Overall, a working model of an interconnected process emerges (persistence → immune/complement activation → endotheliitis/microclots → hypoperfusion/mitochondrial dysfunction); this causal chain is plausible, but its linkage is not yet proven. For targeted therapies, biomarker-based subphenotypes and controlled studies are needed [1][2].

Note: all the mechanisms named above are hypotheses or association findings of varying maturity, not established causes. No concrete self-treatment can be derived from any section. This text serves to convey knowledge, is not a promise of a cure and does not replace medical diagnosis or treatment.